A Battery Lifetime Monitoring and Estimation using Split Learning Algorithm in Smart Mobile Consumer Electronics

The extensive utilization of mobile consumer electronic technology in several areas highlights the importance of maximizing battery energy efficiency. Users may want to ensure a specific duration of battery life to allow continuous use of their smart mobile devices. This research introduces a novel method to tackle the issues associated with micro power batteries in consumer devices by utilising a sensor-based smart IoT system. The study also examines the use of Split Learning, an innovative algorithmic framework that improves the accuracy and effectiveness of state-of-charge monitoring in micro power batteries. This approach reduces power usage and enhances the precision of estimating battery life. The circuits that monitor battery temperature and current remain in a low-power mode, only activating when the controller needs to make measurements, leading to a substantial reduction in power consumption. Integrating Split Learning algorithms is a major advancement in battery monitoring technology, offering exceptional power efficiency and accuracy for upcoming micro power battery applications in consumer devices. The proposed approach accurately predicts the remaining battery charge under various discharge scenarios. The device also includes an automated learning system using Split Learning to safeguard the battery from excessive current consumption, ensuring the longevity and reliability of these crucial power sources. The results of the current research were 0.9340 for testing data and 0.9376 for training data.